TRANSCRIPT 🔗
WEBVTT
00:00:19.373 --> 00:00:24.082
Well, welcome back to Class Act season two, navigating the AI Landscape.
00:00:24.472 --> 00:00:27.713
If you're a leader right now, you probably feeling the tension.
00:00:28.132 --> 00:00:31.823
AI is moving faster than any technology we've ever seen before.
00:00:32.393 --> 00:00:33.533
Every week there's a new model.
00:00:34.343 --> 00:00:38.362
A new agent, a new promise, and the pressure is to act.
00:00:38.576 --> 00:00:41.816
Almost every organization now says the same thing.
00:00:42.326 --> 00:00:44.095
We want to use AI responsibly.
00:00:44.610 --> 00:00:46.011
We care about ethics.
00:00:46.310 --> 00:00:47.750
We want to do it the right way.
00:00:48.456 --> 00:00:53.856
But when you scratch beneath the surface, many leaders struggle to answer a simple question.
00:00:54.276 --> 00:00:59.375
What does responsible AI actually mean when real commercial decisions are on the line?
00:01:00.036 --> 00:01:02.165
That's what today's episode is all about.
00:01:02.735 --> 00:01:14.611
I'm joined by Jeff Watkins, fractional Chief AI Officer and CTO AI advisor, keynote speaker, and one of the UK's most thoughtful voices in AI ethics.
00:01:14.941 --> 00:01:16.441
Security and adoption.
00:01:17.102 --> 00:01:27.572
Jeff has spent more than 25 years building and leading technology at scale, across financial services, healthcare, government, and the private sector.
00:01:27.621 --> 00:01:32.811
By the end of this episode, you'll understand why AI principles matter more than ever.
00:01:33.712 --> 00:01:43.283
How to move from ethical intent to real world execution, and how leaders can set guiding principles that enable innovation rather than slow it, down.
00:01:44.093 --> 00:01:47.123
If you're watching on YouTube, make sure to subscribe.
00:01:47.311 --> 00:01:49.052
Jeff, really proud to have you here.
00:01:49.522 --> 00:01:50.951
Welcome to a Class Act I thank you very much.
00:01:50.971 --> 00:01:52.921
I should hire you as my PM man.
00:01:53.221 --> 00:01:54.811
We've got your wonderful introduction.
00:01:57.031 --> 00:01:57.902
Oh, not at all.
00:01:58.027 --> 00:02:08.556
It's been good, kind, obviously we're, we've just started talking and now and agAIn, and having a few conversations over LinkedIn and stuff, so it's been good to get to know you, so I'm really excited about this one.
00:02:08.887 --> 00:02:09.396
Oh, me too.
00:02:09.997 --> 00:02:22.049
So, so Jeff, listeners who may not know your background, could you give us a quick snapshot of your journey and explAIn how you came to focus so heavily on AI ethics, governance, and responsible adoption?
00:02:22.199 --> 00:02:22.438
Sure.
00:02:23.128 --> 00:02:24.748
I'm a lifetime technologist.
00:02:24.748 --> 00:02:29.133
I've, I started coding when I was six years old on something called the Commodore Vic 20.
00:02:29.399 --> 00:02:31.419
And, dunno how many of your are.
00:02:31.989 --> 00:02:36.459
Listenership will be old enough to remember such a machine had four kilobytes of ram.
00:02:36.459 --> 00:02:39.008
It's pretty primitive and I got the bug.
00:02:39.308 --> 00:02:48.908
I'm one of those very few people in life who are fortunate enough to, from the very beginning, know what they wanted to do with their lives and actually get to do it.
00:02:49.574 --> 00:02:56.123
And I'm sure plenty of people wanted to be fire people, astronauts, something, something, a surgeon and never get to do it.
00:02:56.123 --> 00:03:02.753
Whereas I, I, I, from a very young age decided I wanted to work with technology and managed to be able to do it, which is great.
00:03:03.174 --> 00:03:10.854
So I entered into the, I, I went through the traditional route of going to university back when it was still affordable to do so.
00:03:11.709 --> 00:03:25.199
Doing computer science and then moved into tech consultancy for a company called BGSS in 1999 and BGSS, , was only 23 people then, and grew to be several thousand people and recently sold to CGI.
00:03:25.408 --> 00:03:29.338
I moved between different consults becoming.
00:03:30.179 --> 00:03:35.068
Chief for technology and then chief product and technology officer, and then Chief Technology Officer.
00:03:35.368 --> 00:03:36.568
Different places.
00:03:37.109 --> 00:03:40.769
And, but the biggest thing I think is, is, and the answer.
00:03:40.769 --> 00:03:46.049
So the second part of the question is why, why I, I've come to focus so heavily on things like ethics and governance and responsible adoption.
00:03:46.769 --> 00:03:54.473
It's, I've been through the.com bubble bursting and the boom, I've seen the shift into the cloud.
00:03:55.373 --> 00:03:56.364
But what along the way?
00:03:56.364 --> 00:04:01.313
So I, I'm a technology, so I cared about the technology and that's really all I cared about for the first, you know, decade plus of my career.
00:04:01.313 --> 00:04:08.783
And then I realized how much of an effect technology has on people's lives as the more it gets adopted and that can't be ignored anymore.
00:04:08.813 --> 00:04:18.509
'cause back when it was just my first things were writing, banking, interchange systems that shipped money between banks, they, they didn't really affect end users or people that much.
00:04:19.673 --> 00:04:22.194
But now when you see the high street's completely changed.
00:04:22.283 --> 00:04:25.043
Banking, you know, bank, branches have all but closed down.
00:04:25.463 --> 00:04:29.963
How we interact and how we build relationships is completely, fundamentally changed.
00:04:30.533 --> 00:04:35.274
I remember maybe 10 years ago, 15 years ago, if you sAId you met somebody online, you were a bit weird.
00:04:35.274 --> 00:04:36.413
That's a weird thing to do Now.
00:04:36.413 --> 00:04:38.184
Something like 90% of people meet online.
00:04:38.814 --> 00:04:45.363
So technology and , the web and mobile apps have fundamentally changed our lives and AI.
00:04:46.098 --> 00:04:47.269
Or at least
Load full transcript
this wave of AIr.
00:04:47.269 --> 00:04:53.894
'cause AIr has been going for a very long time, that since the 1950s this could be the is and the reason this unlike something maybe like.
00:04:55.024 --> 00:04:57.244
A blockchAIn, which everybody's banging on about for a few years.
00:04:57.603 --> 00:04:59.043
This is a cultural zeitgeist.
00:04:59.463 --> 00:05:05.463
This is, you know, people now have AI on their phone and they on their tv have it on their fridge and on their kettle.
00:05:06.004 --> 00:05:07.504
It's just a big cultural change.
00:05:07.834 --> 00:05:17.788
And what I think we need to, looking at the, how the web has changed our interactions, and I think AI has more power still to change how we live, is we need to do it in a way that.
00:05:18.338 --> 00:05:20.963
It's human first and not profits first.
00:05:21.324 --> 00:05:24.173
And it needs to support and enhance the human condition.
00:05:24.403 --> 00:05:25.963
For the many, not for the few.
00:05:26.293 --> 00:05:30.694
So that's why we need to think about how we ethically, , implement it.
00:05:30.894 --> 00:05:41.124
So the, the first thing, when you think about the threats of AI harms most people, and this, so I do a lot of stuff on AI safety and we've been doing courses with Blue Impact, which are great if you ever get a chance to do one of the courses.
00:05:41.543 --> 00:05:43.673
And one of 'em is the existential fact, the Terminator.
00:05:44.194 --> 00:05:48.488
Story, but that's not really, that's not really, the problem here, in my opinion.
00:05:48.488 --> 00:05:50.139
At least not, not the, not the closest problem.
00:05:50.139 --> 00:06:16.177
To us, the closest problem is, is people misusing AI either accidentally or on purpose, and especially, , larger technology organisations, That's why I'm so invested in this is because things could go really badly if we don't, and we need voices in this space Now we certAInly do a hundred percent and I think even from my point of view, like working in the kind of data talent market and AI as well.
00:06:17.091 --> 00:06:26.841
Just over the past, I would even say the past month or two, I'm just starting to hear the more a i, ethics, guardrAIls, safety, governance.
00:06:26.841 --> 00:06:31.101
It's starting to come into all conversations, so it's definitely an interesting time.
00:06:31.221 --> 00:06:31.372
Mm-hmm.
00:06:31.627 --> 00:06:34.538
But you have lived through multiple technology waves.
00:06:35.048 --> 00:06:45.487
What, what would you say it is about this moment in AI that makes principles and ethics feel less like a nice to have and more like a kind of leadership imperative?
00:06:45.577 --> 00:06:48.608
So, as I sAId, I think it's the biggest cultural ze guess since the web.
00:06:48.608 --> 00:06:55.177
But, but, but importantly I left a little bit of narrative out there is if you look at the industrial evolutions, they're getting closer together.
00:06:56.168 --> 00:07:05.148
And this, we're arguing now that the argument is now we're in another industrial revolution, , and the other one only started in the mid nineties, whereas previous ones lasted, you know, a hundred plus years.
00:07:05.148 --> 00:07:11.327
So this is compressing time and it, and it's compressing adoption.
00:07:12.137 --> 00:07:21.408
And because everybody will be at some point, forced to adopt AI in some ways, or form or another, or face extinction, it's important that they do it well.
00:07:21.557 --> 00:07:22.728
They do it well, and they do it.
00:07:23.612 --> 00:07:26.552
In ways that supports humans, as in people want to use their services.
00:07:26.822 --> 00:07:31.473
And also in, in line with things like ever moving,, regulation, like the EU AI Act.
00:07:31.562 --> 00:07:34.862
And you know that thanks to the Brussels effect, this will start to spread.
00:07:34.983 --> 00:07:37.983
Now, the EU AI Act is actually quite different to GDPI in many ways.
00:07:38.132 --> 00:07:43.892
GDPR is a lot of management and governance, you know, because it's data handling, governance, and management, whereas.
00:07:45.153 --> 00:07:48.413
The AI Act requires a lot more , on the technical layer to think about it.
00:07:48.622 --> 00:07:56.153
So I was discussing this with somebody recently and they, they, they sort of clued me into this and I was like, all right, that actually makes sense that it's not quite just GDPR V two for AI.
00:07:56.423 --> 00:07:57.793
It's a little bit more, complex than that.
00:07:58.093 --> 00:08:02.783
So if you don't, if you don't have a good handle on this and you can't adopt it, responsibly and ethically.
00:08:03.653 --> 00:08:06.233
It's probably worse than not adopting it at all.
00:08:06.233 --> 00:08:10.432
And you might be left on the, unfortunately, can't consent to the scrap heap of organizational history.
00:08:10.612 --> 00:08:21.442
So I think this is, this is, this is an existential threat from organisations not because of, you know, the Terminator problems, but because of people who don't do it well, will just be forced out of business, in my opinion at least.
00:08:21.713 --> 00:08:22.608
Yeah, and I totally agree.
00:08:23.168 --> 00:08:23.487
I do.
00:08:23.540 --> 00:08:27.990
And almost every organization they say it wants to use AI responsibly.
00:08:28.560 --> 00:08:35.379
In your experience, why do AI principles so often remAIn slideware rather than shaping the real decisions.
00:08:35.830 --> 00:08:41.919
It's like anything when you have values or a vision, , , there's kind of two types of organizations.
00:08:41.919 --> 00:08:43.750
There's ones who just go, oh, we need a values and vision.
00:08:43.750 --> 00:08:49.720
'cause organizations have values and vision and there's those who live in never break character.
00:08:50.590 --> 00:08:54.940
You know, like you get method actors who never break character all the way through filming and like.
00:08:55.450 --> 00:09:02.320
Value, genuine values driven and
vision driven organizations never break character, and they tend to be the ones that are they, they're all or nothing.
00:09:02.320 --> 00:09:06.159
They're either very successful , or unfortunately, they, their value vision just , don't really matter to the world.
00:09:06.164 --> 00:09:07.034
, And they, they die out.
00:09:07.434 --> 00:09:09.325
But they're very much live or die by them.
00:09:09.894 --> 00:09:12.894
So unfortunately, principles the same kind of thing.
00:09:12.985 --> 00:09:14.965
They can be seen as words and words are very cheap.
00:09:14.965 --> 00:09:20.205
You know, nowadays you can put words out there and trade offs are, but the problem is trade offs in organizations are expensive.
00:09:21.524 --> 00:09:21.644
So.
00:09:22.620 --> 00:09:25.470
If they're not properly adopted, then they have no teeth.
00:09:25.649 --> 00:09:28.500
The written as aspirations, but not constrAInts.
00:09:28.740 --> 00:09:32.940
You know, we'd like it to be fAIr and equitable, but then you don't say about in any ways, like, how are you gonna do that?
00:09:33.809 --> 00:09:35.340
How are you gonna make those decisions?
00:09:35.340 --> 00:09:35.909
Reality?
00:09:36.389 --> 00:09:41.850
So if nothing is actually a constrAInt, then nothing changes, you know, nothing ever changes.
00:09:41.850 --> 00:09:42.809
You can't expect it to change.
00:09:43.230 --> 00:09:49.620
And also quite often is if there's nobody owning AI and owning a accountable firm.
00:09:49.801 --> 00:09:52.921
And not only for what AI does, but also how it's implemented.
00:09:53.431 --> 00:09:56.191
So if it's no, if it's just some principles on a page, it's nobody's job.
00:09:56.191 --> 00:09:56.490
No.
00:09:56.610 --> 00:09:58.321
They'll be like, well, it doesn't really apply to me.
00:09:58.321 --> 00:09:59.160
I'll just do what I want.
00:09:59.326 --> 00:10:03.105
And quite often that then means that incentives, incentives fight agAInst them.
00:10:03.285 --> 00:10:16.706
So if you've got people incentivized, people are, are measured like good hearts law, you know, measured agAInst certAIn things If those incentives outweigh the benefits of these principles, then they, they'll all just get left in the dust.
00:10:17.126 --> 00:10:21.777
And this is the same with o, other elements of sort of non-revenue generating things like cybersecurity.
00:10:22.106 --> 00:10:25.106
If the incentive structure is wrong, this is how financial crashes happen.
00:10:25.706 --> 00:10:31.616
'cause if the incentive structures are wrong, people will act to, to meet those, those numbers and then think bad things will happen.
00:10:32.067 --> 00:10:34.407
So I think the difference, right?
00:10:34.557 --> 00:10:41.397
Principles fAIl when they're treated as values alone and they work when they're actually treated as requirements.
00:10:41.966 --> 00:10:48.927
So each principle needs to be turned into a rule, a test, an owner and a gate, rather than just being like an aspirational value.
00:10:49.856 --> 00:10:50.726
Yeah, and I love that.
00:10:51.297 --> 00:10:52.032
Totally agree as well.
00:10:52.101 --> 00:10:58.613
I've I've worked in businesses and had my own business and you know, how cheap it can be when you have these vision days and value days.
00:10:58.613 --> 00:11:05.023
But if you're not really loving them and practicing them and moving them forward, it's, it can be a waste of time.
00:11:05.743 --> 00:11:05.832
Mm-hmm.
00:11:06.163 --> 00:11:06.793
And it's a different level now.
00:11:06.798 --> 00:11:06.908
Mm-hmm.
00:11:07.543 --> 00:11:14.082
So, so when leaders come to you asking for help defining AI principles, what's usually driving the conversations?
00:11:14.743 --> 00:11:18.192
Is it regulation, risk, reputation?
00:11:18.552 --> 00:11:26.582
This is what we found when we first started working when I was at Create Future i'm now doing my own independent,, AI consultancy, which you'll probably shout that out at some point.
00:11:27.532 --> 00:11:37.542
Is we did our first piece of work with, with a very well known name, and it was very, this was early 2023, so it was very successful and we thought everybody would want this, but it turns out.
00:11:38.547 --> 00:11:47.517
The organizations in the uk, especially small to medium sized, and even large organizations in early 2023, were much further left in this journey.
00:11:47.517 --> 00:11:50.846
They had no idea they had very low AI literacy.
00:11:52.256 --> 00:11:58.256
And that meant that all of those things, people were worried about risk, re regulation, reputation, all those kind of things.
00:11:58.706 --> 00:12:00.866
But there also, there's a big fear of missing out.
00:12:00.866 --> 00:12:06.667
But a lot of it's fear driven because they don't understand and they wanted to feel like the, the, they wanted legitimacy.
00:12:07.057 --> 00:12:17.662
They feel like, you know, can we actually adopt AI in a way that feels authentic, that people who adopt and trust, that won't put us at risk of the other of the reputation and regulation, all those kind of things.
00:12:17.871 --> 00:12:29.272
They don't want the blowback and a lot of those lacked an understanding of AI, so very low levels of literacy that then that you need those to feed into.
00:12:29.302 --> 00:12:31.072
I think to start having talks.
00:12:31.447 --> 00:12:37.657
Real serious conversations about AI principles and AI policies and adoption.
00:12:38.886 --> 00:12:47.836
So there's a lot of moving parts that, and the mAIn thing is just because they don't understand that it's like saying, let's engage in as a particular board game or something like that.
00:12:47.836 --> 00:12:48.917
But you've never seen that board game.
00:12:48.917 --> 00:12:50.356
You never, you don't know the rules.
00:12:50.626 --> 00:12:53.986
So you could just go, sure, but I don't know what I'm doing.
00:12:54.177 --> 00:12:57.537
But once you know the rules and engagement, you may not be the best
strategist in it.
00:12:57.537 --> 00:13:01.527
You may not be the ultimate expert in it, but at least you know the rules.
00:13:01.647 --> 00:13:03.356
At least you know the game, you know the board.
00:13:03.626 --> 00:13:07.836
And that's really important, because organizations aren't just about ethics theater.
00:13:08.226 --> 00:13:09.307
They want to know, right?
00:13:09.336 --> 00:13:10.927
How do we capture the upside of AI?
00:13:10.956 --> 00:13:17.167
'cause all the boards want this, but without losing control of the important things that matter to them, the trust and reputation and stuff like that.
00:13:18.547 --> 00:13:20.256
And it's, it's it very interesting.
00:13:20.256 --> 00:13:28.417
You sometimes wonder like some of the organizations that have actually got someone actually owning it in-house, like it's their like, full-time job.
00:13:28.897 --> 00:13:36.197
I've spoke to probably quite a lot of people lately where they've touched on it and they've maybe done it as part of another 10 jobs.
00:13:36.197 --> 00:13:40.687
So it's, I think it's getting to a stage now where it's gonna get much more real.
00:13:41.076 --> 00:13:53.932
I was actually from a running my recruitment point of view, I was actually thinking to actually look at hiring somebody now who specializes just in the subject alone, but because I think it's going to get much bigger.
00:13:54.081 --> 00:13:56.091
Sure, yeah, a hundred percent.
00:13:56.892 --> 00:14:03.636
And you've talked about helping organizations move from AI hype to secure production grade delivery.
00:14:04.177 --> 00:14:10.057
How do strong guiding principles actually enable better outcomes rather than slowing teams down?
00:14:10.057 --> 00:14:12.876
So one of the things I think that, that slows teams down.
00:14:12.966 --> 00:14:22.746
So nowadays, coding , and thanks to actually partially to AI tooling as well, but also things like to do cloud native platforms and SaaS platforms and local platforms.
00:14:23.226 --> 00:14:27.162
Building software is actually not that hard, the code itself.
00:14:28.001 --> 00:14:31.241
The hard bit is building the right thing and building it right?
00:14:32.591 --> 00:14:35.772
So good AI principles, what they do is remove uncertAInty.
00:14:35.772 --> 00:14:37.272
'cause uncertAInty breeds doubt.
00:14:37.272 --> 00:14:40.631
And it means, it, it breeds wrong decisioning, it means regret work.
00:14:40.871 --> 00:14:53.072
So having a nAIled down ethics and, principles that lead into policy and into, basically they end up being decision accelerators.
00:14:53.552 --> 00:15:11.282
They remove them ambiguity, they let teams move forward with a lot of more confidence 'cause instead of having debates about what, about this or that, you know, if you, if you see actually we, we never use per our client's personally identifiable information in this as as like part of the detAIl of the principles.
00:15:11.731 --> 00:15:12.871
Then there's no argument to be had.
00:15:12.932 --> 00:15:15.511
It's been agreed at the exec level and board level.
00:15:15.871 --> 00:15:17.881
We just get on, we, we don't do that.
00:15:18.211 --> 00:15:20.282
So, and that reduces rework.
00:15:20.282 --> 00:15:22.351
It stops people getting stuck at the last mile.
00:15:22.351 --> 00:15:30.631
You know, when you go, when, when security and compliance and governance, people look at your system before you go on live, 'cause you haven't engaged 'em soon enough they go, I'm not sure we can put this live.
00:15:31.081 --> 00:15:33.272
It stops you falling at that, that hurdle.
00:15:33.447 --> 00:15:36.596
Clear constrAInts help people move faster.
00:15:37.037 --> 00:15:37.937
They help people move.
00:15:38.057 --> 00:15:45.226
They help people be more creative 'cause working with, within a universe of no constrAInts, people just, you can just do anything and it becomes bewildering.
00:15:45.647 --> 00:15:48.136
They also help standardize the architecture underneath it all.
00:15:48.136 --> 00:15:51.287
Not just the enterprise architecture, not just a technical architecture.
00:15:51.621 --> 00:15:53.032
So principles do not slow delivery.
00:15:53.032 --> 00:15:53.601
It's quite the opposite.
00:15:53.601 --> 00:16:01.341
It feels paradoxical 'cause you think it would do, it's uncertAInty slows delivery principles replace uncertAInty with like reusable patterns of practice.
00:16:01.761 --> 00:16:02.812
Yeah, no, totally.
00:16:03.111 --> 00:16:06.371
And When you help an organisation define its AI principles?
00:16:06.711 --> 00:16:09.761
What are the core non-negotiables that should always be present?
00:16:10.812 --> 00:16:11.292
Yeah.
00:16:11.292 --> 00:16:13.701
So I did, I did some principle.
00:16:13.731 --> 00:16:19.701
So it's primary research on this across over 50 different organisations internationally, public and private sector.
00:16:20.152 --> 00:16:23.621
And like what are the ones that everybody really cares about?
00:16:24.042 --> 00:16:33.011
And there were five, and this is in order, it was transparency, accountability, fAIrness, privacy, and security.
00:16:33.491 --> 00:16:35.081
So I'll go over those in, in, in order.
00:16:35.741 --> 00:16:39.221
And I think, and this feels like, oh, maybe security should be high.
00:16:39.221 --> 00:16:42.486
But I think actually when it comes to transparency, it's the primary key.
00:16:42.517 --> 00:16:51.096
'cause you people need to know when AI, AI is being used and for what purposes, what's being used for and what it's limits are, but also how secure all the other pieces around it.
00:16:51.096 --> 00:16:55.506
So that transparency, it's like with, when you're adopting agile, transparency is such a primary key to it.
00:16:56.407 --> 00:16:57.966
It's really, really important.
00:16:58.216 --> 00:17:00.647
And that was by far the most important principle.
00:17:01.037 --> 00:17:07.531
There's accountability because there must be a clear AI, you know, owner
for every AI system and decision pathway.
00:17:07.951 --> 00:17:09.392
And that's also really important to people.
00:17:09.511 --> 00:17:14.372
'cause that leads into then like, well what happens when there's fAIrness and things like, you know, that include like harm reduction.
00:17:14.791 --> 00:17:16.771
How do we make sure we're not disadvantaging people?
00:17:16.892 --> 00:17:21.632
How do we actually then if, if you've got fAIrness in place, how do you actually put redress in place if there is something goes wrong?
00:17:21.811 --> 00:17:23.281
And then things that privacy.
00:17:23.686 --> 00:17:25.757
And security a bit more self-explanatory.
00:17:25.757 --> 00:17:32.836
Although privacy in the case, privacy has two different meanings, and I was discussing this with somebody about a year or so ago in cybersecurity.
00:17:32.836 --> 00:17:46.001
It means like keeping stuff, people's stuff kind of con like the cia, a try confidential, but also privacy., In many other terms, it's like the right to your own personal privacy, and that means that maybe things like the right to your data portability, which is actually quite hard with AI.
00:17:46.092 --> 00:17:55.301
If your data's gone into a model, you can't just expunge the data out of a database very easily, picking data, picking learned things out, a model actually quite difficult.
00:17:55.721 --> 00:18:04.616
So I think there were few of those I think that was really important as well things like safety and reliability, making it fit for purpose, really important.
00:18:04.981 --> 00:18:09.267
And, and the, and contestability appeal, override and recourse.
00:18:09.537 --> 00:18:14.396
And these are the, these are the ones I think that you should start with and you can combine some of these together.
00:18:14.576 --> 00:18:28.676
I'd usually say to organizations, don't try and have a list of 20 principles'cause if there's, you know, it's like when, say when everything's important, nothing is, try and keep it to five, but maybe put some of those together, like maybe put privacy and security or security and privacy together.
00:18:28.916 --> 00:18:32.086
Maybe put fAIrness and contestability together.
00:18:32.477 --> 00:18:36.946
Put a couple together so you've got like this, this little list of things that it, you can get it.
00:18:37.307 --> 00:18:42.916
You need to do a little bit of, I think wordsmithing here to make it feel like everybody can pick it and go, right, I understand what we're moving towards here.
00:18:43.156 --> 00:18:45.227
And these principles are our top ones.
00:18:45.227 --> 00:18:54.741
They're not necessarily all of our rules and guardrails, but they're the ones where we're considering things that we should absolutely not in any way compromise on.
00:18:55.491 --> 00:18:56.422
Yeah, absolutely.
00:18:56.632 --> 00:18:57.352
Oh, thanks for that.
00:18:58.251 --> 00:19:12.707
And, and how do leaders balance speed V safety, innovation, V control, mm, commercial pressure V versus ethical responsibility without ending up with the principles that are either meaningless or paralyzed?
00:19:12.826 --> 00:19:12.916
Mm-hmm.
00:19:13.787 --> 00:19:14.297
So.
00:19:15.467 --> 00:19:19.307
I think then you need to think about how you actually turn everything into a risk-based operating model.
00:19:19.396 --> 00:19:30.017
So it's not just a set of inspirational words, so you can use bounded speed, so it's not move fast versus being safe, it defines where we can move quickly.
00:19:30.832 --> 00:19:34.811
So imagine and the AI, the AI Act kind of outline something similar to this.
00:19:34.811 --> 00:19:37.092
Basically, think of it almost like a traffic light model.
00:19:37.211 --> 00:19:50.801
There are low risk internal use cases, or even some low risk, external use cases that are, , like, say like a recommendation system, but internally, maybe it's internal drafting things, summarization of nonsensitive content, lightweight controls and glare.
00:19:50.981 --> 00:19:52.902
They're green kits and you can just move fast with those.
00:19:53.422 --> 00:19:54.922
You've got very simple controls in place.
00:19:55.132 --> 00:20:07.241
There's amba stuff where there's more something that's a little bit more customer facing, like customer facing chat bots where it's maybe the sense, the use case is a little bit more, has more risk involved, decision support systems.
00:20:07.541 --> 00:20:19.211
So then you need a little bit, then you need to think about, you can't just go fast and break things with decision support 'cause if you're turning people down for I don't know, uh, loans or other financial things that, that could be really damaging.
00:20:19.211 --> 00:20:21.132
So you don't wanna move fast and break things with that.
00:20:21.672 --> 00:20:28.362
Then you've got the red stuff, and that could be stuff that, and the EU Act has kind of got one of these layers, like you must never.
00:20:29.231 --> 00:20:37.092
But there'd be some stuff where you might do it in, in the red zone, but you'd probably wanna strongly consider not doing it or at least putting loads of control around it.
00:20:38.126 --> 00:20:43.807
Be that things like hiring, using AI for hiring, , medical decisioning, anything with safeguarding.
00:20:43.807 --> 00:20:53.176
So if you are writing out a chat bot for, students who are under the age of 18, there's a strong elements of safeguarding, then you need to really think about how that's implemented and put in place.
00:20:53.477 --> 00:20:55.426
So there's got to be lots of formal sign off for that.
00:20:55.596 --> 00:20:55.807
Yeah.
00:20:55.807 --> 00:20:58.422
And the question is, do you do this at all with AI Frankly?
00:20:59.136 --> 00:21:03.487
Now, there may be in the future it might be, it might be might, you might just be too ahead of your time a little bit.
00:21:03.981 --> 00:21:08.676
And, and the another thing is right, is start to pre-approve patterns, not projects.
00:21:08.676 --> 00:21:13.896
If you're doing this every single time for every, like going from first principle, every single time, of course it's gonna slow down.
00:21:14.721 --> 00:21:20.422
But once you've got the patterns, the primitives of the building blocks of things, like here's the things we can do, here's the types of things we can do.
00:21:20.662 --> 00:21:31.011
If you are going to take use case A and use case B is very similar to use case A, there's no extra risk and you've already passed use case A and it's already effectively in production Use case B should be simple for you.
00:21:31.011 --> 00:21:33.346
You shouldn't have to go through the whole rigor roll agAIn.
00:21:34.017 --> 00:21:43.507
If you first time you put something into life, if you're putting minor changes into life or something, you wouldn't necessarily need to go through the whole life process you should be able to, you know, do continuous delivery.
00:21:44.287 --> 00:21:51.297
Same kind of thing with AI patterns is think about how you actually reuse these things rather than going from, from first principles.
00:21:51.576 --> 00:22:02.902
But also the align, align everybody's incentives around this, making sure that people are all the areas of the delivery life cycle are thinking with the same hymn sheet, and have very, very fast escalation paths.
00:22:02.981 --> 00:22:16.092
Really you have to be quite strict on outcomes and harms, but flexible on some of the methods around, you know, principles should create the guardrails that let teams deliver much faster and not become like, a maze that they kind of have to escape like rats really.
00:22:16.571 --> 00:22:16.781
Yeah.
00:22:16.781 --> 00:22:16.872
No.
00:22:16.872 --> 00:22:17.442
Fascinating.
00:22:18.192 --> 00:22:20.711
This will be really good for a lot of the people listening to the show.
00:22:21.201 --> 00:22:21.682
Definitely.
00:22:21.682 --> 00:22:26.307
Is there any AI principles you don't trust when you see them written down?
00:22:26.576 --> 00:22:32.636
Because they sound good, but are vague, performative, or impossible to operationalize?
00:22:33.266 --> 00:22:39.236
The devil's in the detail So there's a few principle shaped sentences that read really nice and do almost nothing.
00:22:39.445 --> 00:22:41.246
You know, we all use AI ethically.
00:22:42.145 --> 00:22:43.256
Whose ethics is that?
00:22:44.050 --> 00:22:44.980
Measured by how?
00:22:45.181 --> 00:22:47.431
First of all, ethics is never a checklist.
00:22:47.790 --> 00:22:54.961
Having done AI ethics certifications and courses is one of the first things to teach is that you can't just have a, a checklist of AI Ethics.
00:22:55.320 --> 00:23:01.141
Ethics are systems and they're complex, and they're also quite different between organizations.
00:23:01.300 --> 00:23:14.011
This is why some people wouldn't choose to work in certAIn industries, that doesn't mean that there's, there's a universal 'cause there are no universal objective ethics Really, which is frustrating, but so yeah.
00:23:14.082 --> 00:23:16.241
We will be transparent, like, okay, transparent.
00:23:16.241 --> 00:23:18.672
To whom is that to the board?
00:23:18.672 --> 00:23:19.751
Is that to regulators?
00:23:19.751 --> 00:23:20.622
Is that to users?
00:23:21.179 --> 00:23:25.739
We will avoid bias and ensure fairness and bias free is actually impossible entirely.
00:23:25.828 --> 00:23:27.868
And it's very, it's almost impossible.
00:23:27.868 --> 00:23:29.669
It's, and it's impractical entirely for your bias.
00:23:29.669 --> 00:23:34.679
Like you need to figure out what fairness and bias criteria you're going to try and minimize and, and how to prove that.
00:23:35.078 --> 00:23:44.034
There's a few of these, but if you can't turn a principle to some kind of observable, falsifiable test, some kind of no go, no, go.
00:23:44.334 --> 00:23:47.003
Then it's probably just a slogan on the website that looks really nice.
00:23:47.634 --> 00:23:50.304
So I'm not agAInst words, the words of any of these.
00:23:50.624 --> 00:23:52.814
I'm agAInst them being unqualified.
00:23:53.834 --> 00:23:56.443
And the principle has to have some kind of cost you're willing to pay.
00:23:56.713 --> 00:23:58.153
They need to be clear and explicit.
00:23:58.513 --> 00:24:01.574
Don't make them something really vague that you can hide behind.
00:24:01.634 --> 00:24:09.223
You know, even if you do have a top level one word, principle of transparency, underneath that needs to be some kind of rolldown where you actually qualify that.
00:24:09.943 --> 00:24:10.064
No.
00:24:10.153 --> 00:24:10.483
Totally.
00:24:11.233 --> 00:24:18.374
And one of the hardest challenges is translating principles into behavior for engineers, product teams, and data scientists.
00:24:18.824 --> 00:24:22.544
What does good implementation actually look like at a team level?
00:24:22.548 --> 00:24:38.173
Yeah, I mean, that, that's, I think that's an interesting one because one thing I do really take a lot of heart from it is, and when I first joined the industry, people in, in product engineering, we didn't even call it product engineering, then largely didn't understand ethics or care about 'em.
00:24:38.638 --> 00:24:41.128
Because, 'cause for the most part, they weren't touching users that much.
00:24:41.608 --> 00:24:47.278
I'd say most people nowadays, or a good chunk of people nowadays, are much more ethically tuned in than they were.
00:24:47.805 --> 00:24:55.650
So at the team level, how do you get this, first of all, make sure everyone understands the risk tiers, you know, these traffic light kind of system.
00:24:55.920 --> 00:25:04.375
So people know we're working on something high risk here, so we need to, we need to really think more about it versus this is the lowest possible risk internal use case.
00:25:04.855 --> 00:25:06.984
We don't need to worry too much about this.
00:25:07.015 --> 00:25:08.335
We still have some guardrails.
00:25:08.875 --> 00:25:19.865
But building these guardrails really, not just guidance, actually have guardrails that you can review agAInst and go, We're not gonna ship this, but also try to build against secure patterns and, and good patterns as well.
00:25:19.894 --> 00:25:27.305
You know, if you, if you're building lots of rag systems, make sure you've got a pattern catalog and reusable things that mean you don't have to think about it afresh every time.
00:25:27.454 --> 00:25:29.974
People shouldn't have to spend, people should think about the ethics.
00:25:29.974 --> 00:25:37.694
You should think about the, they should be thinking about , the consequences of the AI they're building, but really they want to think about the business value they're delivering.
00:25:38.365 --> 00:25:48.474
So try and make that simple by again, having primitives that are packaged up that you can reuse, , but also making sure there's ownership and oversight, making sure people are involved.
00:25:48.654 --> 00:25:55.345
Shift left on, lemme talk about shift left, security shift, left this shift, left that quality, ethics, security all should be shift left.
00:25:55.525 --> 00:25:57.630
Make sure there's ownership, making sure there's real.
00:25:58.204 --> 00:26:05.075
Human oversight, but then also thinking about how you monitor that in production as well, because it's not just, it's, it's not a one and done.
00:26:05.075 --> 00:26:07.295
It's not, it's not one, one and done thing in a delivery.
00:26:07.954 --> 00:26:10.805
Systems are living, breathing things, especially AI based systems.
00:26:11.375 --> 00:26:16.795
I mean, when responsible AI's done well, you've put it into primitives and and genuine guardrails and tests.
00:26:17.035 --> 00:26:26.634
It should be as invisible as possible because it's baked into all your workflow, your platforms, your templates, your libraries, and the teams don't have to necessarily remember all the principles.
00:26:26.634 --> 00:26:29.184
You know, they should be all kind of part of the system.
00:26:29.457 --> 00:26:52.856
And could you share an example, anonymized if needed, where a clear ethical stance, a guiding principle, generally changed a technical or product decision yeah, and this, this, this is about automation and it was where we, we, we are having a good conversation about how far, how much do we want to automate certain tasks away, and how much do we want and how much do we want the human to look?
00:26:52.933 --> 00:26:59.023
the sensitivity of what they're doing was hard decision on this one hard, like very clear statement.
00:26:59.023 --> 00:27:05.473
We will not be going to replace any people with AI and all important decisioning will be done with a human in the loop.
00:27:07.169 --> 00:27:08.669
So that shapes the solution.
00:27:08.848 --> 00:27:14.038
That means there's no agentic system automatically turning down a particular case or thing.
00:27:15.659 --> 00:27:30.838
That's a hard decision now that in the future may be somewhat, limiting, some might say, but then you think, well, all of the other bits of the AI automation that helped reduce that workload on that team were very, you know, the mAIn thing was, is the, very busy team, very busy workload.
00:27:32.068 --> 00:27:39.209
The other things you can implement reduce 80% of that work or a good chunk of that work, that means they can make a better choice when it comes to the decisioning process.
00:27:39.209 --> 00:27:40.199
They don't have to rush that.
00:27:40.638 --> 00:27:46.219
Imagine if you've got better systems and tools and documentation that may allow you to make better decisions.
00:27:46.429 --> 00:27:53.134
'cause some of the other drudge work that is is actually humans are not as good at the more inaccurate.
00:27:53.763 --> 00:27:56.493
It's taking their, their, their sucking the air out the room.
00:27:56.644 --> 00:28:00.993
And actually what they need to do is make this quite sensitive decision based on evidence.
00:28:02.013 --> 00:28:05.794
So it's better for the better for the people, it's better for the customer.
00:28:06.094 --> 00:28:06.929
Yeah, no, I totally agree.
00:28:07.729 --> 00:28:08.838
A good, great example.
00:28:08.838 --> 00:28:09.288
Thanks.
00:28:09.949 --> 00:28:13.909
And you've led large scale AI literacy and upskilling programs.
00:28:14.449 --> 00:28:21.648
How important is AI literacy in making ethical principles stick, especially for non-technical leaders and boards?
00:28:22.663 --> 00:28:32.723
So, I can't even stress how incredibly important is because it, agAIn, going back to the game example, how do you govern or make any kind of strategy in a game you have no understanding of.
00:28:33.923 --> 00:28:38.304
And that means that some organizations go they just get the tech team to start using AI 'cause they don't understand the risks.
00:28:38.693 --> 00:28:42.743
They want the benefits of AI without the groundwork, which is great up until an incident happens.
00:28:43.193 --> 00:28:45.743
And there's also lots of organizations that are far too fearful.
00:28:45.743 --> 00:28:48.804
And I think in the UK we're, we're quite a conservative bunch.
00:28:48.804 --> 00:28:51.864
And I don't mean that politically conserv, it means in we're we're quite reserved.
00:28:51.864 --> 00:28:59.213
And we would quite, a lot of organizations, especially SMEs are, are too fearful of it 'cause they don't understand how to constrain the risks of using AI.
00:29:00.233 --> 00:29:07.064
So there's a very, there's, they're so afraid of it that they're willing to forego the benefits of using modern AI and that could eventually destroy the business.
00:29:07.334 --> 00:29:14.953
In, in reality, the lack of AI literacy at the board and exec level will usually lead to AI getting stuck in pilot hell not delivering the business benefits expected.
00:29:15.344 --> 00:29:20.443
So as with have a lot of other things, a lot of this modeling has to come from top
down.
00:29:20.594 --> 00:29:23.743
Don't just leave AI to tech people'cause AI is a cultural phenomenon.
00:29:24.334 --> 00:29:36.084
, And it's not just the mechanisms of AI, it's the systems around it and it's the methods of use that are cultural zeitgeist and that the board and the exec absolutely need to understand this and set the hand on the teller and start modeling good behavior.
00:29:36.848 --> 00:29:49.648
That's fascinating times it is, and sitting at the intersection of AI security and governance, where the organization most often underestimate risk when deploying generative and agentic AI?
00:29:52.233 --> 00:30:16.413
Oh, this is a big question just because of the, so I was recently doing a talk at, at a large security group and I was worried that, 'cause he was talking to AI security, specifically LLM security, and I was worried that I'd be roundly whipped out the room because it was, I was telling them, you know, everything they already knew and I was, I asked the people in the room about 200 people, how many people had a good handle on LLM security and not a single hand went up.
00:30:16.413 --> 00:30:16.624
So.
00:30:17.679 --> 00:30:23.318
Security consents are not well understood even by, security professionals in this field, frankly.
00:30:24.219 --> 00:30:30.058
So wider organizations, even less so now then we start talking about agent AI.
00:30:30.598 --> 00:30:35.398
Once AI breaks out the chat window, the threats kind of multiply.
00:30:35.608 --> 00:30:43.548
They have access to file systems, APIs, databases, payment services, all kinds of things that are quite worrying you, you know.
00:30:44.344 --> 00:30:47.854
And the attacks agAInst these systems are novel and sometimes not very intuitive.
00:30:47.943 --> 00:30:50.733
And sometimes it's not even just attacks, it's just stuff going wrong.
00:30:51.453 --> 00:30:53.943
Because when we, it's not just agents anymore.
00:30:53.943 --> 00:30:58.683
It won't just be agents soon, it'll be flocks of agents, it'll be collaborative groups of agents.
00:30:58.804 --> 00:31:08.074
It'll be effectively organizations of agents who will act in emergent ways, not even necessarily because of attack, just like that attack in emergent ways.
00:31:09.574 --> 00:31:23.409
So people need to understand how to threat model this quite urgently, I'd say, before you start playing with it too hard and really doing some very big damage to your organization or potentially running our business.
00:31:23.709 --> 00:31:30.308
We need, this is one thing I do with people is help them understand how to threat model or actually one of the threat modeling with them.
00:31:31.104 --> 00:31:34.733
How to underst understand how to threat model these kinds of systems.
00:31:34.733 --> 00:31:45.653
'cause it's a very diff so there's traditional security still exists and, and still matters in these cases, but there's a whole new world of security concerns that you need to think about how to model 'em.
00:31:46.584 --> 00:31:52.193
No, you get worried even listen, you talk like that and I'm not even kind of running a in a large scale enterprise.
00:31:53.453 --> 00:31:55.733
You could spend hours just on this topic.
00:31:56.423 --> 00:32:00.203
It's, there's a lot to it, but I, yeah, I would, but it would make it far too long An episode.
00:32:01.614 --> 00:32:02.453
Oh, maybe another one.
00:32:02.453 --> 00:32:03.054
I'm looking forward.
00:32:03.233 --> 00:32:04.554
I'm looking forward to sharing this.
00:32:04.554 --> 00:32:15.773
It's, yeah, it's just when you're talking, you realize how important it is because accountability is a huge issue when an AI system causes harm or unintended consequences.
00:32:16.104 --> 00:32:18.443
Who actually owns that decision in practice?
00:32:21.429 --> 00:32:25.209
Oh, this is, this is a philosophical question as much as it is a technical question.
00:32:25.459 --> 00:32:29.963
You know, under our current laws and understanding of what AI is and isn't.
00:32:30.354 --> 00:32:31.344
It can't be the AI.
00:32:31.344 --> 00:32:39.144
Let's be really clear, there's a really good, , book called Robot Rules by Jacob Turner, and it talks about there's, there's not a set of laws for horses.
00:32:39.413 --> 00:32:44.534
So when a horse tramples somebody, the horse isn't legally responsible as as a legally responsible entity.
00:32:44.743 --> 00:32:47.653
They are an entity in, in, after the fact, just in case.
00:32:47.804 --> 00:32:52.544
Just in the same way their house is an entity, but they're not legally responsible for their actions.
00:32:53.443 --> 00:33:05.354
So the risk is, is not, is, is, is really, is is not having an accountability structure inside your organization 'cause the organization will be then found responsible and that will find its way up to the CEO or something like that.
00:33:05.864 --> 00:33:10.483
And so it shouldn't, so it'll come as a surprise when something bad happens and the fingers start pointing.
00:33:10.814 --> 00:33:15.284
And, the best thing is to have accountable under sport AI.
00:33:15.854 --> 00:33:23.844
So you can't just make the developers themselves of the systems accountable because a lot of these risks in AI, well, some of those might be, security risks.
00:33:23.844 --> 00:33:37.564
Some of them might be just, just the, the organizations taking an unacceptable risk in how they're leveraging AI and you can't hold the developers necessarily responsible for that 'cause they don't necessarily have the background in that particular field to understand what they've done.
00:33:38.104 --> 00:33:40.263
You know, it could, sometimes it could be the third parties.
00:33:40.263 --> 00:33:47.403
Imagine if the you are using open AI's APIs and those APIs do something really bad 'cause of open AI's oversight, what they're doing.
00:33:47.753 --> 00:33:48.953
That could be the third party.
00:33:49.243 --> 00:33:51.913
Or it could be even be the end user if they're misusing
AI.
00:33:52.034 --> 00:33:58.183
So that we OpenAI a while ago found various unfriendly nation states using their services.
00:33:58.459 --> 00:34:01.519
For bad uses of AI and, and we're blocking it.
00:34:02.058 --> 00:34:06.169
But as the implementer, mostly it's going to be you, the accountability, frankly.
00:34:06.439 --> 00:34:12.619
So you should have named people and you should have enough, you should be doing due diligence on suppliers to make sure you're not using dodgy ones.
00:34:12.619 --> 00:34:13.759
And where, where's the data going?
00:34:13.759 --> 00:34:15.079
What's the data being used for?
00:34:15.498 --> 00:34:20.298
You know, have you put enough protections in place to, to prevent misuses of real AI services?
00:34:20.539 --> 00:34:37.833
So really it's who owns that decision is you get to choose and it has to be somebody, and it might, it probably be more than one person, and it should, and those people should be speaking to legal and to user groups and to the security groups, et cetera.
00:34:38.623 --> 00:34:43.889
If the system can actually affect people, it needs a named person who can press the stop button.
00:34:44.039 --> 00:34:47.248
And if you do not have that facility.
00:34:48.164 --> 00:34:51.193
You have a, a potential big problem waiting for you.
00:34:51.494 --> 00:34:53.054
But the thing is, it shouldn't be hard to get there.
00:34:53.083 --> 00:34:56.173
That that's one of the, I think one of the easier things to solve, to be honest with you.
00:34:57.039 --> 00:35:03.809
look, with regulation evolving rapidly, how should leaders think about designing AI principles that are future proof?
00:35:04.409 --> 00:35:05.878
Not just compliant today.
00:35:07.168 --> 00:35:07.528
Hmm.
00:35:08.248 --> 00:35:11.458
Yeah, I think people, people should be thinking about this now.
00:35:11.728 --> 00:35:16.378
And I as to say, start with something that's something that's reasonably robust.
00:35:16.719 --> 00:35:27.778
Now principles are not outcomes and then not laws, but I'd start with things like looking at the EU AI Act to get a good understanding of it and also take inspiration from organizations who are known to be using AI ethically.
00:35:28.139 --> 00:35:31.588
But principles are outcomes you want, and they're not references.
00:35:31.889 --> 00:35:32.969
'cause laws do change.
00:35:32.969 --> 00:35:34.378
So the EU AI Act will change.
00:35:34.648 --> 00:35:47.559
Outcomes don't, you know, so, you know, people, people being able to contest out, decisions survives any micro changes in legislation , and you should anchor all of these to risk and not technologies.
00:35:47.739 --> 00:35:51.818
If I, anything, all of in all of this conversation, technology is the last thing you're talking about.
00:35:52.059 --> 00:35:59.059
Imagine it's an alien technology and the actual, doesn't matter if it's on AWS or on Microsoft, it's all about management, management of risk.
00:35:59.699 --> 00:36:03.389
And design these, design these things for auditability by default.
00:36:03.389 --> 00:36:07.619
Make sure you can actually test these things so you can test them, and you can get the results of those tests.
00:36:07.623 --> 00:36:13.778
And you, and if you ask the question, you know, where are our principles at , do you feel like they're working for us?
00:36:14.079 --> 00:36:15.429
Can you prove that we're sticking to 'em?
00:36:15.429 --> 00:36:19.599
If you can't answer any of these questions and you've got a big problem, and think about making them as evergreen.
00:36:19.809 --> 00:36:21.429
Try and make them as evergreen as possible.
00:36:21.608 --> 00:36:27.039
Don't make them too specific on individual things or likely to change.
00:36:27.293 --> 00:36:32.813
It is like with core values, if you have core values, you shouldn't be changing your core values every week.
00:36:32.914 --> 00:36:37.204
That means that you don't have robust core values that are evergreen.
00:36:37.324 --> 00:36:39.844
Cause like compliance, any given time is a snapshot of time.
00:36:39.844 --> 00:36:45.313
You know, it it because, and the world moves, because future proofing these is, is a kind of a capability.
00:36:45.733 --> 00:36:53.923
So you think about the ability to evidence your intent, test for harm and adapt safely, you know as the rules and the technology both change underneath you.
00:36:54.344 --> 00:36:54.583
Yeah.
00:36:55.003 --> 00:36:55.664
I can imagine.
00:36:55.664 --> 00:36:59.954
Over, over the years, you've, you have sat in many boardrooms as well, Jeff.
00:36:59.983 --> 00:37:00.193
Mm-hmm.
00:37:00.474 --> 00:37:08.574
, When you're in a boardroom, what signals tells you, like a leadership team generally understand AI ethics versus simply repeating the right language?
00:37:09.204 --> 00:37:10.434
I think there's a few things.
00:37:10.434 --> 00:37:20.588
First of all, that they know about the named accountability and it exists, that they understand the accountability challenges,, they're asking for direct evidence.
00:37:21.248 --> 00:37:24.369
They understand enough about it to ask for direct evidence of things.
00:37:24.619 --> 00:37:31.818
, They can name what the, what the, the, the, the yeses and the nos are when it comes to goes and no GOs and they say, no, we don't, we definitely don't do that.
00:37:31.849 --> 00:37:36.289
If they can emphatically answer those things, they probably have a reasonable grasp of them.
00:37:36.708 --> 00:37:39.829
And they're talking concrete scenarios and not abstractions.
00:37:40.128 --> 00:37:44.668
You know, they think about, here's exactly what could go wrong in our business context.
00:37:45.179 --> 00:37:47.309
And what we detect and how we shut it down.
00:37:47.518 --> 00:37:54.208
So really there's a few, there's a few things that when people talk in concrete terms, it probably means they've got a good handle on it.
00:37:54.599 --> 00:37:56.518
You know, there there's performative stuff you
sell.
00:37:56.518 --> 00:37:58.768
We're, we're fully committed to something, something, something, something.
00:37:59.068 --> 00:38:03.199
It's funny, it's like, you know, when people say to be honest and then they lie to your face.
00:38:04.039 --> 00:38:08.088
When people say, we're fully committed to, sometimes it's go, are you committed to this at all?
00:38:08.298 --> 00:38:09.438
Do you even care about this?
00:38:09.539 --> 00:38:21.148
I think of just using policy as a crutch and just saying, well, our policy say such and such, without any further rational thought behind it, is usually a good sign that they're not really taking it, it seriously, it's quite performative.
00:38:21.648 --> 00:38:41.498
Ethicsand maturity I think it's all really visible when leaders can describe the trade offs, you know, and when you've worked close to the lines, 'cause it's where the lines are in the ethics that it's hard, you know, if you're close to a line, that's where it gets really interesting and where you choose fAIrness over speed or revenue and things like that, you know, and how you actually prove that that worked.
00:38:41.648 --> 00:38:47.570
And if a board only asked one or two questions about AI principles, what would you say those questions should be?
00:38:48.260 --> 00:38:56.079
I think the only asked two, and I think there's, a couple that make the entire system of their thinking reveal itself.
00:38:56.079 --> 00:38:58.389
And it's like, first of all, it's where are our red lines.
00:38:59.409 --> 00:39:02.829
And who has the authority to stop in the case of us hitting a red line.
00:39:03.159 --> 00:39:08.710
You know, if they can't answer specifically on that, you know, it's just a bit performative.
00:39:09.110 --> 00:39:16.789
And then the second one is, you know, was it, oh, people always quote we demming saying, in God we trust all those must bring evidence or data.
00:39:16.849 --> 00:39:20.539
And Deming never sAId that, which is amusing because, 'cause it's based on evidence and data.
00:39:20.539 --> 00:39:22.000
But, same in this, you know.
00:39:22.570 --> 00:39:25.760
In God we trust all others, when it comes to ethics, must bring evidence.
00:39:26.090 --> 00:39:29.389
You know, how do we evidence compliance with our principles in real operations?
00:39:29.389 --> 00:39:31.369
If they can't answer that, then it's, it's theater.
00:39:31.760 --> 00:39:41.420
If we could squeeze in the third one, it's like, you tell me what's the most likely way AI could harm your customers or your staff or your business in our context.
00:39:41.869 --> 00:39:43.849
And if they have no idea, then haven't thought about it enough.
00:39:43.849 --> 00:39:45.469
Can you take, get off of it quick?
00:39:45.949 --> 00:39:46.280
Yeah.
00:39:46.579 --> 00:39:46.909
Really do.
00:39:47.690 --> 00:39:48.139
They do.
00:39:48.349 --> 00:39:50.269
We're coming close to the end here Mm-hmm.
00:39:50.269 --> 00:39:50.280
Hmm.
00:39:50.420 --> 00:39:55.579
So, for a leader listening, who knows, they should define AI principles but hasn't started yet.
00:39:56.119 --> 00:39:59.780
What is the first practical step they should take in the next 30 days?
00:40:00.710 --> 00:40:03.800
Well, I, I'd get them to contact me and hire my services.
00:40:04.219 --> 00:40:04.510
Yeah, of course, of course.
00:40:04.510 --> 00:40:05.570
I think that would be a good start.
00:40:05.639 --> 00:40:09.599
But think about what, what would I suggest, like is do a focused AI reality check.
00:40:09.965 --> 00:40:14.795
Doing some in some workshops tend to one, like a, one, like a one page planner page effectively, you know?
00:40:15.105 --> 00:40:18.284
And really what, what you things, well you, first of all, you need somebody who's accountable.
00:40:18.284 --> 00:40:19.844
You need some kind of accountable sponsor.
00:40:20.175 --> 00:40:20.445
Yeah.
00:40:20.864 --> 00:40:28.485
You need to also know what the heck's going on already 'cause was it over 85% of organizations have got people in them using generative AI for day-to-day work.
00:40:28.815 --> 00:40:30.764
Only 40% have got actual licenses.
00:40:31.485 --> 00:40:34.695
Which says the, the, the great AI gap is massive.
00:40:34.965 --> 00:40:37.994
So you need to understand what is actually happening in your organization.
00:40:37.994 --> 00:40:40.934
I cannot stress that enough 'cause it's, and it's like a bring out your dead.
00:40:40.985 --> 00:40:42.034
It's an amnesty.
00:40:43.054 --> 00:40:44.644
If you've been doing it, you're not gonna get shot.
00:40:45.284 --> 00:40:45.885
Within reason.
00:40:46.304 --> 00:40:47.710
Just you need to tell us.
00:40:47.739 --> 00:40:51.639
We need to find out what is going on and there'll be, and there'll be fragmentation, there'll be cost savings all over the place.
00:40:51.880 --> 00:40:53.614
Pick some use cases that you need to think about.
00:40:54.309 --> 00:40:54.670
Yeah.
00:40:55.000 --> 00:41:02.230
And then think about what, then think about ideating that what, what the trade offs and the risks and what data is involved, who's affected, how much is gonna cost?
00:41:02.260 --> 00:41:05.110
And then start, how do we build out these principles?
00:41:05.110 --> 00:41:07.929
Which ones do you have to like with the values you have to battle test them another way.
00:41:07.929 --> 00:41:09.550
Which of these ones are we interested in?
00:41:09.789 --> 00:41:14.769
And take your top five principle statements in really plain English, no jargon.
00:41:15.835 --> 00:41:28.795
Or whatever your native language is really plain And what are the non-negotiable rules behind them and who are the accountable people owns them and what is, and then also like what's your risk tiering model for stuff as well?
00:41:28.974 --> 00:41:31.465
'cause you'll know in your own organization what's the low high risk.
00:41:32.094 --> 00:41:43.304
So don't start with some fancy manifesto and, and words really start with like your real use cases, the red lines, the owners, and publish your V 0.1 paper that that sets you in motion.
00:41:44.429 --> 00:41:45.480
Ah, great advice.
00:41:45.659 --> 00:41:46.199
Great advice.
00:41:46.530 --> 00:41:56.219
And finally, if you could leave leaders with one mindset shift about AI ethics, some thing that would generally change how they approach AI decisions, what would it be?
00:41:56.639 --> 00:42:03.034
Stop a treating AI ethics as a, oh, we do this to be nice to, to make, you know, maybe it makes us look good as a B core, something like that.
00:42:03.389 --> 00:42:05.875
And start treating it as like critical systems system safety.
00:42:06.195 --> 00:42:09.014
AI ethics isn't some hand wavy statement of intent.
00:42:09.375 --> 00:42:21.465
It's risk engineering for human impact of your users, your staff, and the world, frankly, you know, when people start thinking of it as safety critical, and absolutely critical for the business, rather than being oh, it's these nice, nice things that we put on our webpage.
00:42:21.465 --> 00:42:26.295
You know, they start thinking about in more in security terms, like, what could go wrong and how will we know when it goes wrong?
00:42:26.715 --> 00:42:31.065
They'll start to funding this of the actual testing, the monitoring and things like that an instant response.
00:42:32.219 --> 00:42:34.619
We have really strict red lines and things like that.
00:42:34.619 --> 00:42:45.409
You know, if it matters enough to you to actually use AI and automate it, it should matter enough for you to make it safety critical, you know, safety critical system in, in that, in that particular safety critical tier.
00:42:45.409 --> 00:42:48.019
You know, even the green tier needs to be secure.
00:42:49.744 --> 00:42:52.715
You don't get, people always say, we don't need a Rolls Royce solution.
00:42:52.894 --> 00:42:59.135
Yeah, you can build a car that doesn't have, you know, a nice wall, a dash and everything, but you can't build cars nowadays without seat belts.
00:42:59.405 --> 00:43:00.125
You know, it's not legal.
00:43:00.184 --> 00:43:12.619
So cut your cloth to your, to your risk level, but always make sure even the, even the green, even the green light scenarios still need basic levels of security and safety and, and, and ethics behind them.
00:43:12.889 --> 00:43:14.780
I've genuinely learned a lot here myself.
00:43:14.780 --> 00:43:20.929
It's a subject I don't know loads about, so I've, I've done more listening today, so it's been really, really interesting for me.
00:43:20.929 --> 00:43:25.639
So I know, I know the people who listen to the podcast will get a lot from this as well.
00:43:26.239 --> 00:43:26.750
Definitely.
00:43:26.809 --> 00:43:27.380
Really good.
00:43:27.650 --> 00:43:34.250
Hey, Jeff, just before we go, can you just let people listening are watching know where to find you and how they can get in touch with you.
00:43:34.610 --> 00:43:35.059
Sure.
00:43:35.059 --> 00:43:35.389
Okay.
00:43:35.389 --> 00:43:37.099
So, you can find me on LinkedIn.
00:43:37.380 --> 00:43:38.369
I'm Jeff Watkins.
00:43:38.489 --> 00:43:49.545
For those listening, you won't know what I look like, but hopefully you'll see the thumbnails , I am just opening now so by the time this podcast goes out, you'll probably fully launched a, a small business called North Star Intelligence.
00:43:49.885 --> 00:43:51.775
And that is AI consultancy.
00:43:51.864 --> 00:43:57.355
That's what I'll be doing most of my time fractional pieces, not, it doesn't necessarily need to be huge statements of work.
00:43:57.355 --> 00:44:05.755
It'll be working with a number of different industries on AI strategy, AI ethics, AI policy, governance, and also the technical elements of it.
00:44:05.755 --> 00:44:09.034
I've just finished my AI master's degree, which is really fun.
00:44:09.215 --> 00:44:10.864
So the technical elements as well.
00:44:11.074 --> 00:44:15.394
You can also find me on the Compromising Positions Podcast, which I run with my partner, Leanne Potter.
00:44:15.724 --> 00:44:18.664
If you're searching for that podcast, it's on all your normal podcast platforms.
00:44:18.875 --> 00:44:26.605
Search for compromising positions, podcast not compromising positions unless you, especially not on a work laptop, unless you wanna have the security team asking you some.
00:44:27.105 --> 00:44:36.585
Awkward questions, you know, and reach out to me, reach out to me, listen to the podcast, follow my business when it launches on LinkedIn and just have a, have a, have a conversation.
00:44:36.954 --> 00:44:37.795
I look forward to it.
00:44:38.545 --> 00:44:44.635
Ah, brilliant, Jeff, I look forward to continuing these conversations with you and see where you can take your business as well.
00:44:44.965 --> 00:44:45.969
Hugely successful.
00:44:45.969 --> 00:44:48.940
Okay, well thanks very much and yeah, thank you.
00:44:48.940 --> 00:44:48.960
Thank you.
00:44:49.059 --> 00:44:50.170
And I'll speak to you soon.
00:44:50.744 --> 00:44:51.295
Cheer Jeff.
00:44:51.295 --> 00:44:51.454
Thanks.
00:44:51.605 --> 00:44:52.094
Bye-Bye.